This proposal has the long-term objective of treating purine nucleoside phosphorylase deficiency (PNP) by gene transfer. The research employs dogs which, because of a virtual absence of red cell PNP, are exceptionally suitable as in vivo models of PNP deficiency. PNP deficiency is associated with defective T cell and normal B cell immunity. Patients invariably die unless treated with some form of enzyme replacement. Gene transfer into pluripotent marrow stem cells is an attractive protocol. Also, as tissue specific expression of PNP is not required for therapy, gene transfer into autologous skin fibroblasts represents an effective alternative route for the delivery of therapeutic genes. To achieve these goals, human PNP cDNA will be inserted into retroviruses containing dominant-acting selectable genes. The specific aims are: 1) Infection of mouse fibroblasts with different PNP-virus constructs to determine the one most efficiently expressing PNP. 2) Study of transduction and expression of human PNP cDNA into human PNP- B LCLs and skin fibroblasts. PNP expression will be monitored and purine metabolism studied in phenotypically corrected cells. 3) Infection of dog skin fibroblasts in vitro. Study techniques for transplantation of infected autologous canine skin fibroblasts in vivo. Monitor proliferation, human PNP expression and effects on purine metabolism. 4) Canine marrow will be infected in vitro and PNP expression monitored in hematopoietic progenitor cells. 5) Transplant infected autologous marrow into dogs and monitor PNP expression in hematopoietic progenitor cells. This research provides a model of somatic cell gene transfer in large animals that will be applicable to several enzyme deficiency diseases.